If you are a cable TV system engineer, you can't be blamed for feeling just a little bit smug these days. It wasn't that long ago that everywhere you looked, national publications such as Forbes, Wired, The Wall Street Journal, etc. were pushing and shoving to be the first to hang the cable industry out to dry. Strident headlines asked, "Where are the new cable services?" "What about the return path?" "Cable promises; where are they?" And so on.

The much-hyped HFC network, it seems, was at the heart of the problem. "Noisy and unreliable." "Costs billions of dollars to upgrade to two-way." Even, "Outdated analog technology."

Figure 1: CTB vs. power level.

Yet, somehow, through all this, the only really useful (and affordable) Internet access is being provided by the cable industry. Cable companies are beginning to offer telephone service, and, guess what — subscribers are switching! The HFC network is emerging as the best high-speed access network there is. And, partly as a result, cable equity values are soaring.

Makes us all feel good, doesn't it? On the other hand, we know that the HFC network is far from perfect. Ingress can be a problem, although strategies are emerging to control the problem. Problems persist getting large-scale networks to work properly, but they can be fixed with good engineering and good training. And we all know that the passing of time takes a toll on outside plant. Networks that work just fine a month after turn-on develop problems down the road.

Never mind, we say. Because today's hybrid fiber/coax network design, with 500- to 1,000-home nodes, is only the beginning. Waiting in the wings is the world's most powerful, most flexible, and most cost-effective high-speed, two-way network. Namely, HFC Fiber-to-the-Last Active (HFC-FTLA).

There is little doubt that taking linear fiber optics deep into the network, down to the 50- to 100-home level, will reduce or even eliminate operational problems that still need to be controlled with today's designs. Reduced ingress, easier maintenance and fewer active devices in the network all produce immediate benefits to the operator.

There are three reasons why FTLA may emerge sooner rather than later:

  1. Increased focus on cable telephone service (AT&T acquires TCI) will put the spotlight even more on network reliability.
  2. Emergence of new amplifier technology will create greater reach from nodes and line amplifiers.
  3. New optical technology will enable the extension of fiber past existing nodes.
Growing focus on cable telephone

The proposed AT&T acquisition of TCI brings the issue of cable telephony to the forefront. While most cable operators currently focus on high-speed access to the Internet, the deployment of standard voice telephone services over HFC networks continues to grow worldwide. If cable telephony is to make a credible alternative to services offered by existing local exchange carriers (LECs), reliability and voice quality must be equal to or better than that available over copper twisted pair lines. Furthermore, MSOs may find that if they partner with long distance carriers like AT&T (or others), the penetration of telephone services will exceed their expectations. High penetration of telephone service and high-speed Internet access will put pressure on the capacity of today's networks, especially as the distribution plant ages, and maintenance becomes a critical issue.

These two factors, substantial penetration of service and increased focus on reliability, call in to question the suitability of existing node sizes. MSOs may find it prudent and cost-effective to reduce the number of homes served by a single node to well below the 500-home level. If they do, they will find that recent developments in RF and optical technology will make their job easier.

New amplifier technology

After decades of deploying coaxial networks based on silicon amplifiers, the cable industry has begun to see the impact of new materials and new devices on the performance characteristics of line amplifiers. GaAs MESFET (Metal Semiconductor Field Effect Transistor)-based amplifiers are already offered by equipment manufacturers. These amplifiers provide improved distortion, lower noise and higher gain than equivalent amplifiers designed with silicon devices. The results have been quite dramatic. Lower distortion, higher gain and lower noise all promise reduced network cost by reducing the number of actives per mile.

And now, new amplifiers using heterojunction devices have already been demonstrated that surpass the GaAs results*. Compared to the GaAs MESFET devices, heterojunction transistors are advanced designs, although the advantages of the heterojunction have been known for several decades. Only in the last decade have advances in materials processing enabled researchers to manufacture the heterojunction device in volume.

Consider the performance characteristics shown in Figure 1. Here, two prototype amplifiers constructed with heterojunction transistors are compared against a standard 870 MHz silicon bipolar power hybrid. It is easy to see the advantages of the heterojunction devices in higher output power for a given level of CTB performance. At a CTB level of approximately 64 dB, the GaAs amplifier has an output power advantage of more than 5 dB, compared to the silicon amplifier. And, this performance improvement is obtained at 30 percent lower DC power consumption.

Every 1 dB increase of output power from an amplifier will, ideally, increase the number of homes served from that amplifier by the same amount, or about 26 percent. This is never achieved in practice, because some power is lost in the additional coaxial transmission path needed to reach these additional homes. In low-density networks, the extra "reach" might only be five percent, while in dense networks, the number of new homes served might be 15–20 percent.

Suppose for a moment that an "average" network could reach 12 percent more homes per amplifier for each 1 dB increase in amplifier output. Then, a 5 dB increase in output power would result in approximately 75 percent more homes reached from a single line extender. This has a dramatic impact on HFC network design. The number of actives per mile of plant would drop from the current value of five to something less than three. In a 500-home node, which today might utilize about 25 amplifiers, the number would drop to 14 or 15.

An additional advantage of passive coax is that all the third-order distortion may be allocated to the final stage in the receiver. Thus, we can expect an additional 3 dB to 4 dB output power from the amplifier, which would reduce the number of RF amplifiers within an average 500-home node to fewer than 10.

Consider the improvements as shown in Table 1. The key to these improvements, ironically, is not new fiber optic technology, but new RF technology.

Improved receiver design

The operational advantages of heterojunction devices can also be applied to optical receivers. Using these heterojunction transistors' noise advantages, experts expect that cable TV receiver noise can be significantly reduced from the 6–7 pA/(check)Hz of today's receivers. Lower receiver noise translates to lower optical input power to the receiver.

Supporting deeper fiber

Pushing fiber past the node of today's networks poses some optical technology challenges. First of all, some means must be found for increasing the amount of optical power available at the node to support splitting the fiber to new receivers downstream. Secondly, when the demand for dedicated bandwidth outstrips the spectrum available to the original 500-home node, some means must be found to introduce new dedicated signal spectrum without installing new fibers back to the headend.

In reality, the transition to FTLA architectures is not likely to happen overnight. The first step would probably be to push the optical fiber past the existing node to perhaps the 125-home level, with perhaps two additional line extenders downstream. To provide the additional power for this extension, there are several possible strategies. In many cases, the transmitter at the headend has more than enough power to serve a single node, and the signal is split to feed two or more receivers. By installing new transmitters of the same performance (often at a fraction of the original price for the existing transmitter), the signal can now be split at the node and transmitted to the 125-home level. This is shown in Figure 2.

The second step would be to extend the optical fiber to the final active device before the subscriber. Again, more power is needed, perhaps an additional 2 dB or 3 dB. The potential sources for the extra power are: (1) a higher-power 1310 nm DFB transmitter, (2) a medium-power 1550 nm transmitter and Erbium doped fiber amplifier (EDFA), or (3) a 1310 nm optical amplifier at the headend, if such technology becomes commercially feasible.

The last stage would be to begin providing narrowcast signals to the individual receivers. To this point, the assumption is that all subscribers within the original 500-home territory are still receiving the same signal set, and that the bandwidth demands of individual subscribers can still be met with the 750 MHz (or 860 MHz) spectrum on the cable. Someday, however, the demand for bandwidth may exceed even that. When that day arrives, new 1550 nm dense wavelength division multiplexing (DWDM) transmitters can be installed at the headend and signals transmitted over a single fiber to the original node site, where they will be de-multiplexed and combined with the remaining broadcast channels for delivery to individual receivers.

Figure 2: Providing broadcast and narrowcast signals to the level of 125 homes can be accomplished with a DWDM technology overlay.
Can we afford it?

The primary disadvantage of reducing the node size is cost. More optical receivers, more return path transmitters and more fiber mean higher cost per home passed. Calculating that cost is beyond the scope of this article, but in light of expectations for additional revenue, as penetration of new services grows, the cost of FTLA will not be out of line with current HFC upgrade costs and pay-back times.

We have shown that today's 500-home node with approximately 25 line extenders could be replaced by approximately 10 optical receivers utilizing heterojunction amplifier technology. Initially, these could be served by a single AM+QAM transmitter, expanding to multiple narrowcast QAM transmitters as demand for bandwidth increased. Detailed cost calculations of these networks are better left to equipment manufacturers, not to technology suppliers, but the authors would claim that initial deployments of FTLA will be within range of today's network costs, and will pay immediate dividends in reliability and reduced maintenance.


The key to advanced wired communications networks is fiber optics. As fiber pushes closer to the subscribers' premises, signal quality and reliability will increase, while operating costs will decrease. Recent advances in optical and RF technology have made HFC Fiber-to-the-Last-Active networks even closer to reality, at a time when the focus on reliability and bandwidth is becoming even more relentless.


Network technologyNumber of amplifiers in 500 home node
Standard design25
New amplifier technology15
Lower CTB specifications10
Table 1: Number of amplifiers needed in 500-home node with new amplifier technology advances.

Reference: Chris Day, Proceedings of the Emerging Technologies Conference.